And What That Means for the Future of Physical AI

As robots move from factories into homes, sidewalks, and offices, tech companies are facing a new challenge: it’s not just about performance anymore. It’s about perception.

Across major cities, delivery robots now navigate sidewalks with expressive “eyes,” rounded bodies, and even names. When one such robot struggled through floodwaters in Los Angeles last week, social media users didn’t mock it. They felt sorry for it.

This shift is not accidental. It’s engineered.

Designing for Acceptance, Not Just Functionality

Historically, robots operated in controlled industrial environments, used by trained technicians. Today, AI-powered machines increasingly interact directly with everyday people.

According to human-robot interaction researchers like Ellie Sanoubari, this changes everything. A robot designed for a warehouse can prioritize efficiency. A robot operating on a public sidewalk cannot afford to feel threatening.

Design elements such as:

Larger heads
Big, circular eyes
Rounded shapes
Soft sounds and gestures

are intentionally chosen because they trigger deeply rooted human responses associated with infants and pets. In other words, cuteness lowers resistance.

For companies building consumer-facing robotics, human acceptance is now a product requirement.

Case Study: DoorDash’s Dot

DoorDash designed its autonomous delivery robot, Dot, with this philosophy in mind.

Dot is capable of navigating urban environments at speeds up to 25 mph. But its defining features are not mechanical. They are psychological:

A rounded body, because humans prefer curves over sharp angles
Large circular eyes that indicate direction
Eye contact behavior to signal pedestrians when to cross
Audible cues to announce arrival

These design choices are not cosmetic. They reduce friction in real-world deployment. Trust becomes a scalability factor.

For robotics companies, “human acceptance engineering” is becoming as important as autonomy itself.

From Utility to Personality

Another example comes from Interaction Labs, which built an interactive robot lamp called Ongo. To shape its personality, the company even collaborated with Alec Sokolow, known for his work on Toy Story.

Ongo behaves like a hybrid between:

A digital assistant
A pet
A concierge

It speaks in a cartoon-like voice, reacts physically to interactions, and builds familiarity over time.

The strategic shift here is subtle but important. These devices are no longer just tools. They are becoming characters.

And that has implications.

The Emotional Dependency Risk

As AI gains a physical presence, the psychological impact intensifies.

Researchers warn that the same emotional attachment risks seen with chatbots may become stronger when embodied in robots. This is particularly sensitive in:

Children’s environments
Elderly care
Emotional support robotics

Companies developing these systems must balance engagement with transparency. Users should understand clearly that they are interacting with a machine.

Without that clarity, the line between assistant and companion becomes blurred.

The “Uncanny Valley” Constraint

Sunday Robotics is building Memo, a household robot designed to load dishwashers and fold laundry.

Its aesthetic intentionally avoids hyper-realism. Developers aim to stay out of the “uncanny valley,” the psychological discomfort humans experience when something looks almost human, but not quite.

Memo’s appearance has been compared to Baymax from Big Hero 6: clearly non-human, but emotionally approachable.

This design balance is delicate:

Too mechanical → feels cold and unapproachable
Too human → feels unsettling
Too toy-like → undermines capability

Robotics companies now operate at the intersection of industrial design, psychology, and AI engineering.

The Bigger Trend: The Decade of Physical AI

At events like the Consumer Electronics Show hosted by the Consumer Technology Association, robotics presence has grown rapidly. Hundreds of robotics exhibitors now showcase AI-powered hardware, including robotic pets designed for emotional support.

Industry analysts argue that the 2020s began as the “intelligence decade,” focused primarily on software and generative AI. The second half of the decade may be defined by something different:

Physical AI.

The software capabilities are maturing rapidly. Hardware is catching up. And as robots leave labs and enter shared human spaces, their success will depend on more than computational power.

They will depend on: